Data redundancy remote multi-activity method and system based on industrial Internet of Things

Abstract

The invention provides a data redundancy remote multi-activity method and system based on an industrial internet of things. The method comprises the following steps: S1, collecting data to a nearest edge node in real time; s2, identifying and enhancing key data in the data, and caching or delaying common data in the synchronous data; and S3, transmitting the key data to a cloud node, further detecting and judging whether an early warning signal appears or not, if so, increasing the number of copies of the key data or shortening the synchronization time, and if not, maintaining the original data redundancy strategy and synchronization period, and continuing to detect. According to the method, the AI prediction module is deployed at the edge node, the redundancy strategy is adaptively stored in a layered manner and dynamically selected, and a pluggable consistency protocol and a fault self-healing mechanism are combined, so that multi-place multi-activity and high-reliability data management in an industrial Internet of Things scene is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial Internet of Things, and in particular, relates to a data redundancy remote multi-active method and system based on the industrial Internet of Things. Background Art

[0002] In today's industrial field, data processing technology based on the Industrial Internet of Things is becoming increasingly important. As an important aspect of it, data redundancy and multi-site active methods and systems aim to improve data security, reliability and availability.

[0003] However, in practical applications, the off-site storage and synchronization of data is difficult and requires solutions to network latency, data consistency, and other issues. Since the devices in the Industrial Internet of Things are widely distributed and the amount of data is huge, how to ensure the fast and accurate transmission and synchronization of data between different locations is an urgent problem to be solved.

[0004] Secondly, the complexity of the system also brings challenges to maintenance and management. To achieve data redundancy and multi-site active-active, multiple links and technologies are involved, including data backup, recovery, and fault switching, which makes the system architecture and management more complicated, increasing the difficulty and cost of operation and maintenance.

[0005] Patent document CN118316981B discloses a cloud-edge collaborative system for data processing of IoT devices, which includes a cloud module and multiple edge node modules, wherein: the edge node module includes: a data acquisition submodule; a data preprocessing submodule; a local storage submodule; a data transmission submodule; a response submodule; a self-recovery submodule; an autonomous decision-making submodule; the cloud module includes: a cloud receiving submodule; a cloud analysis submodule; a cloud storage submodule; a cloud control submodule; the cloud module and the edge node module are connected through a network to achieve two-way data transmission and control instruction transmission. This solution ensures that the system can automatically isolate the problem area, restart the service and verify the recovery effect when a failure occurs, improves the stability and reliability of the system, and reduces manual intervention and downtime. However, it cannot solve the problem of excessive occupancy of core storage and network bandwidth and difficulty in controlling overall operating costs. This problem needs to be solved urgently. Summary of the invention

[0006] In view of the defects in the prior art, the object of the present invention is to provide a data redundancy remote multi-active method and system based on industrial Internet of Things.

[0007] According to the present invention, a data redundancy remote multi-active method based on industrial Internet of Things includes: Step S1: collect data to the nearest edge node in real time; Step S2: Identify and enhance key data in the data, and cache or delay common data in the synchronization data; Step S3: The key data is transmitted to the cloud node, and then detected and determined whether a warning signal appears. If the result is yes, the number of copies of the key data is increased or the synchronization time is shortened. If the result is no, the original data redundancy strategy and synchronization cycle are maintained, and detection continues.

[0008] Preferably, in step S1, the data includes: key data and common data; the key data is real-time fault alarm information of the production line and preset information confirmed based on equipment operation status; the common data is operation log and status record.

[0009] Preferably, in step S2, the process of identifying the key data is to select data and determine whether the value of the selected data exceeds a preset threshold. If the result is yes, the selected data is classified as key data; if the result is no, the selected data is classified as ordinary data; In the step S2, the key data is enhanced by an AI prediction module; the AI ​​prediction module is used to analyze input data and generate prediction results; The key data is composed of features; the mathematical expression of the features is:

[0010] in, Indicates Features represents the normalization function for feature extraction; Indicates Key data, Indicates the first time offset; Indicates Key data, Indicates the second time offset; In step S2, the AI ​​prediction module is used for data prediction, and the mathematical expression of the data prediction is: =σ(W +b) in, represents the probability of failure, W represents the weight, b represents the bias vector, and σ represents the activation function; the activation function is Sigmoid or Softmax; Based on the data prediction, the optimization of the fault switching time of the AI ​​prediction module is expressed as follows:

[0011] in, Indicates the optimization of failover time, The time for the prediction model to respond in advance. is the data redundancy time, The failover time.

[0012] Preferably, in step S3, the early warning signal is a warning of high risk or load surge; The number of copies of the key data is 2; The duration of the synchronization cycle is 30 seconds; The process of judging the high risk or load surge is to judge whether the current CPU or memory usage is greater than 80%, or the data throughput per unit time is greater than 10MB / s, or the failure probability score of the AI ​​prediction model of the edge node is greater than 0.75. If the result is yes, a high risk or load surge warning is issued, and the number of copies of the key data is increased to 3~5 copies, and the synchronization cycle is shortened to 5~10 seconds; if the result is no, the original data redundancy strategy and synchronization cycle are maintained, and detection continues.

[0013] According to the present invention, a data redundancy remote multi-active system based on industrial Internet of Things includes: Module M1: collects data to the nearest edge node in real time; Module M2: Identify and enhance key data in the data, cache or delay the common data in the synchronization data; Module M3: Transmit the key data to the cloud node, and then detect and determine whether a warning signal appears. If the result is yes, increase the number of copies of the key data or shorten the synchronization time. If the result is no, maintain the original data redundancy strategy and synchronization cycle, and continue to detect.

[0014] Preferably, in the module M1, the data includes: key data and common data; the key data is real-time fault alarm information of the production line and preset information confirmed based on the equipment operation status; the common data is operation log and status record.

[0015] Preferably, in the module M2, the process of identifying the key data is to select data and determine whether the value of the selected data exceeds a preset threshold. If the result is yes, the selected data is classified as key data; if the result is no, the selected data is classified as ordinary data; In the module M2, the key data is enhanced by an AI prediction module; the AI ​​prediction module is used to analyze input data and generate prediction results; The key data is composed of features; the mathematical expression of the features is:

[0016] in, Indicates Features represents the normalization function for feature extraction; Indicates Key data, Indicates the first time offset; Indicates Key data, Indicates the second time offset; In the module M2, the AI ​​prediction module is used for data prediction, and the mathematical expression of the data prediction is: =σ(W +b) in, represents the probability of failure, W represents the weight, b represents the bias vector, and σ represents the activation function; the activation function is Sigmoid or Softmax; Based on the data prediction, the optimization of the fault switching time of the AI ​​prediction module is expressed as follows:

[0017] in, Indicates the optimization of failover time, The time for the prediction model to respond in advance. is the data redundancy time, The failover time.

[0018] Preferably, in the module M3, the early warning signal is a warning of high risk or load surge; The number of copies of the key data is 2; The duration of the synchronization cycle is 30 seconds; The process of judging the high risk or load surge is to judge whether the current CPU or memory usage is greater than 80%, or the data throughput per unit time is greater than 10MB / s, or the failure probability score of the AI ​​prediction model of the edge node is greater than 0.75. If the result is yes, a high risk or load surge warning is issued, and the number of copies of the key data is increased to 3~5 copies, and the synchronization cycle is shortened to 5~10 seconds; if the result is no, the original data redundancy strategy and synchronization cycle are maintained, and detection continues.

[0019] According to a computer-readable storage medium storing a computer program provided by the present invention, when the computer program is executed by a processor, the steps of the data redundancy remote multi-active method based on the industrial Internet of Things are implemented. An electronic device provided according to the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the data redundancy remote multi-active method based on the industrial Internet of Things are implemented.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts edge AI prediction algorithm to enhance key data redundancy and ensure fault switching speed, thereby solving the problem of untimely response of systems deployed in multiple locations when dealing with emergencies.

[0021] 2. The present invention solves the problem of excessive occupancy of core storage and network bandwidth and difficulty in controlling overall operating costs by adopting a layered strategy of "multiple active critical data + edge caching of non-critical data".

[0022] 3. The present invention solves the problem of difficulty in achieving comprehensive protection of industrial Internet of Things data at the transmission and storage levels in multi-location scenarios by adopting a unified security strategy of zero trust and cross-regional access control.

[0023] 4. The present invention solves the problem of insufficient system scalability and reusability in different industrial scenarios and equipment scales by adopting pluggable protocols and conflict resolution modules. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings: Figure 1 A schematic diagram of the implementation process of the multi-site deployment and AI prediction redundancy mechanism provided by the present invention; Figure 2 A schematic diagram of the implementation process of tiered storage and cross-region multi-active synchronization provided by the present invention; Figure 3 A schematic diagram of the self-healing switching process under fault detection and zero-trust security provided by the present invention. DETAILED DESCRIPTION

[0025] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0026] The present invention deploys AI prediction modules at edge nodes, adaptively hierarchically stores and dynamically selects redundancy strategies, and combines pluggable consistency protocols and fault self-healing mechanisms to achieve multi-site multi-active and highly reliable data management in industrial IoT scenarios. The solution takes into account both the strong consistency of key data and the final consistency of massive data, and ensures full-link encryption and refined access control under the zero-trust security framework.

[0027] The present invention combines edge computing, cloud computing and intelligent scheduling strategies to achieve efficient, low-latency data redundancy and multi-site collaborative active operation, namely Active-Active. The details are as follows: The present invention utilizes configurable redundancy strategies, namely replica replication, erasure coding, and tiered storage, to ensure that core business data can be quickly restored under different failure scenarios. The present invention uses machine learning models to perform predictive analysis on sensor data, identify the priority of key data, dynamically adjust the redundancy strength and synchronization frequency, and improve network and storage utilization efficiency. The present invention adopts flexible consistency protocols, such as Raft and Paxos, and pluggable conflict resolution strategies, taking into account the differentiated requirements of strong consistency scenarios, such as control instructions, and eventual consistency scenarios, such as non-critical logs. The present invention introduces a distributed health management module to automatically identify and isolate anomalies such as node failures and network congestion, and redirect traffic and reallocate resources through intelligent scheduling. In the industrial Internet of Things environment, the present invention provides end-to-end encrypted transmission, role access control, and identity authentication based on a zero-trust architecture to meet compliance and security requirements in multiple locations. The innovative points of the present invention are as follows: Based on dimensions such as data importance, business priority, and regional compliance requirements, we dynamically select replica replication or erasure coding, or back up key data in multiple locations. For low-value data, we use edge local cache + cloud merging to reduce overall storage costs. We integrate a lightweight machine learning engine on edge nodes to predict potential failures or network congestion by analyzing historical and real-time data. When high load or failure risk is predicted, we can trigger the adjustment of the number of replicas, synchronization frequency, or routing strategy in advance to prioritize synchronization of key data to secure and stable nodes. For core operations that require strict consistency, such as production control instructions, we use strong consistency protocols, namely Raft / Paxos guarantees. For massive logs or sensor data, we use vector clocks or eventual consistency, supplemented by configurable conflict resolution strategies such as timestamp priority and business rule merging, to reduce network bandwidth and synchronization overhead. In addition to common heartbeat detection, the system comprehensively evaluates indicators such as network latency, resource utilization, and node processing latency, and triggers migration or switching in advance when abnormal trends are detected. After the failed node is restored, data compensation and automatic merging can be performed based on incremental logs and version information to reduce operation and maintenance intervention. Through the zero-trust network framework, authentication and continuous security assessment are performed on access to edge nodes, cloud nodes, and internal components. A multi-layer access control model is introduced and combined with industrial protocol features to prevent unauthorized access and tampering of data. The multi-layer access control model includes MAC / RBAC / ABAC. Specifically, the present invention uses an edge AI prediction algorithm; The system collects key data in real time , and perform preprocessing and feature extraction to obtain feature vectors. ={ }, where the mathematical expression of feature extraction is:

[0028] in, Indicates Features, i.e. the result of feature extraction; represents the normalization function for feature extraction; Indicates Key data, including: Indicates the time offset; The normalization function of the feature extraction is used to smooth and normalize the time series data; The key data is used to construct a feature set; The time offset is used to smooth and normalize the time series data;

[0029] in, Indicates the length of the time window; Represents the index of a specific time point within the time window, which is used to traverse each data point, i.e. ; The length of the time window is the time interval used to calculate the feature in the time series; Based on the results of feature extraction, using historical data D hist Training the prediction model of the AI ​​prediction module M , and its prediction formula is: =σ(W +b) in, represents the probability of failure, W represents the weight, b represents the bias vector, and σ represents the activation function; the activation function is Sigmoid or Softmax; By optimizing failover time based on predictions, , the mathematical expression is:

[0030] in, The time for the prediction model to respond in advance. is the data redundancy time, is the fault switching time. Through the edge AI algorithm, the system can Trigger data redundancy and switching in advance to significantly improve redundancy efficiency and switching speed, thereby ensuring key data and system stability.

[0031] The specific implementation process is as follows: Multi-site deployment and AI prediction redundancy mechanism implementation process, such as Figure 1 As shown; Step A1: Collect data to the nearest edge node in real time; Step A2: Instruct the AI ​​prediction module of the edge node to identify and enhance key data according to the characteristics and working conditions of the data; cache or delay synchronization of ordinary data; The AI ​​prediction module is a computing module built based on an artificial intelligence algorithm, which is used to analyze input data in real time and generate prediction results for future states.

[0032] Step A3: transmitting the key data to the cloud node; Step A4: Detect and determine whether there is a high risk or a surge in load. If yes, increase the number of copies of the key data or shorten the synchronization time. If no, maintain the original data redundancy strategy and synchronization cycle and continue regular monitoring.

[0033] In step A2, the key data is real-time fault alarm information of the production line or high-priority data that has a direct impact on the operation of the equipment. The identification method or basis is the edge node AI prediction model or predefined threshold, such as temperature and pressure that are automatically identified as key data when they exceed the limit; In step A2, the ordinary data is low-priority data such as daily equipment operation logs and regularly collected status records, which do not require immediate processing or can be synchronized in batches at a lower frequency.

[0034] In step A4, high risk and load surge are specifically judged as follows: when the CPU or memory usage rate exceeds 80% continuously, or the data throughput per unit time exceeds 10MB / s, or the failure probability score of the edge node AI prediction model is greater than 0.75, it is judged that high risk or load surge occurs. At this time, the number of copies of key data is increased from the original 2 copies to 3-5 copies, and the synchronization cycle is shortened from the original 30 seconds to 5-10 seconds. If any of the above conditions is not met, the original redundancy strategy and synchronization cycle remain unchanged.

[0035] In other words, it looks like this: Industrial equipment or sensors transmit the collected real-time data to the nearest edge node, such as A and B. The AI ​​prediction module in the edge node identifies which data is critical data based on the equipment operating conditions and data characteristics, and which data needs to be synchronized first and enhanced for redundancy. Ordinary data can be cached locally at the edge node or synchronized with a delay. After verification, the critical data is transmitted to the cloud node C1, which monitors the network availability and system load in real time. If the test results are normal, the data is automatically synchronized to other multi-active nodes to ensure data consistency and high system availability. The AI ​​module continuously monitors the health indicators of each node. Once high risks or a sharp increase in load are predicted, the number of copies of critical data will be automatically increased or the synchronization interval will be shortened to ensure data availability and rapid recovery capabilities. The implementation process of tiered storage and cross-regional multi-active synchronization, such as Figure 2 As shown; Based on AI predictions or preset rules, edge nodes divide data into "critical data", i.e. high priority, and "ordinary data", i.e. low priority, and adopt different storage and synchronization strategies for each. For high-priority data, the system will use multiple copies or erasure codes for strong redundancy, and perform real-time or quasi-real-time synchronization between nodes across regions to avoid data loss due to single-point or single-area failures. Edge nodes can cache ordinary data locally, or synchronize it to the cloud at a lower frequency to reduce bandwidth and storage costs. A multi-active architecture is formed between cloud nodes and between cloud and edge nodes, which has disaster tolerance, fault switching and rapid recovery capabilities. The AI ​​module can make collaborative predictions among multiple nodes and allocate resources in the best way. Fault detection and self-healing switching process under zero-trust security, such as Figure 3 As shown; The system judges the health of the node by integrating multiple indicators, namely heartbeat detection, network delay, AI model prediction, and CPU / memory load. Once an abnormal trend or sudden failure is identified, the disaster recovery switching process is immediately entered. During the fault switching process, the system verifies the security status of the remaining available nodes, that is, dynamic identity authentication under the zero-trust concept, to avoid switching traffic to unavailable nodes or abnormal nodes. Available nodes take over business requests and continue to provide complete services to the edge or cloud to ensure production continuity in industrial scenarios. After the faulty node is repaired, it will be reconnected to the system under the zero-trust security framework, and the update differences during the fault period will be compensated by incremental synchronization or data merging, and finally return to the multi-active cluster. The present invention also provides a data redundancy off-site multi-active system based on the industrial Internet of Things, which can be implemented by executing the process steps of the data redundancy off-site multi-active method based on the industrial Internet of Things, that is, those skilled in the art can understand the data redundancy off-site multi-active method based on the industrial Internet of Things as a preferred implementation of the data redundancy off-site multi-active system based on the industrial Internet of Things.

[0036] Those skilled in the art know that, in addition to realizing the system and its various devices, modules, and units provided by the present invention in a purely computer-readable program code, it is entirely possible to realize the same functions in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered as a hardware component, and the devices, modules, and units included therein for realizing various functions can also be regarded as structures within the hardware component; the devices, modules, and units for realizing various functions can also be regarded as both software modules for realizing the method and structures within the hardware component.

[0037] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A data redundancy remote multi-active method based on industrial Internet of Things, characterized in that: include: Step S1: collect data to the nearest edge node in real time; Step S2: Identify and enhance key data in the data, and cache or delay common data in the synchronization data; Step S3: The key data is transmitted to the cloud node, and then detected and determined whether a warning signal appears. If the result is yes, the number of copies of the key data is increased or the synchronization time is shortened. If the result is no, the original data redundancy strategy and synchronization cycle are maintained, and detection continues.

2. The data redundancy remote multi-active method based on industrial Internet of Things according to claim 1 is characterized in that: In step S1, the data includes: key data and common data; the key data is real-time fault alarm information of the production line and preset information confirmed based on the equipment operation status; the common data is operation log and status record.

3. The data redundancy remote multi-active method based on industrial Internet of Things according to claim 2 is characterized in that: In step S2, the process of identifying the key data is to select data and determine whether the value of the selected data exceeds a preset threshold. If the result is yes, the selected data is classified as key data; if the result is no, the selected data is classified as ordinary data; In the step S2, the key data is enhanced by an AI prediction module; the AI ​​prediction module is used to analyze input data and generate prediction results; The key data is composed of features; the mathematical expression of the features is: in, Indicates Features represents the normalization function for feature extraction; Indicates Key data, Indicates the first time offset; Indicates Key data, Indicates the second time offset; In step S2, the AI ​​prediction module is used for data prediction, and the mathematical expression of the data prediction is: =σ(W +b) in, represents the probability of failure, W represents the weight, b represents the bias vector, and σ represents the activation function; the activation function is Sigmoid or Softmax; Based on the data prediction, the optimization of the fault switching time of the AI ​​prediction module is expressed as follows: in, Indicates the optimization of failover time, The time for the prediction model to respond in advance. is the data redundancy time, The failover time.

4. The data redundancy remote multi-active method based on industrial Internet of Things according to claim 1 is characterized in that: In step S3, the early warning signal is a warning of high risk or sudden load increase; The process of judging the high risk or load surge is to judge whether the current CPU or memory usage is greater than 80%, the data throughput per unit time is greater than 10MB / s, or the failure probability score of the AI ​​prediction model of the edge node is greater than 0.

75. If the result is yes, a high risk or load surge warning is issued, and the number of copies of the key data is increased to 3-5 copies, and the synchronization cycle is shortened to 5-10 seconds; If the result is no, the original data redundancy strategy and synchronization period are maintained, and detection continues.

5. A data redundancy remote multi-active system based on industrial Internet of Things, characterized in that: include: Module M1: collects data to the nearest edge node in real time; Module M2: Identify and enhance key data in the data, cache or delay the common data in the synchronization data; Module M3: Transmit the key data to the cloud node, and then detect and determine whether a warning signal appears. If the result is yes, increase the number of copies of the key data or shorten the synchronization time. If the result is no, maintain the original data redundancy strategy and synchronization cycle, and continue to detect.

6. The data redundancy remote multi-active system based on industrial Internet of Things according to claim 5 is characterized in that: In the module M1, the data includes: key data and common data; the key data is the real-time fault alarm information of the production line and the preset information confirmed based on the equipment operation status; the common data is the operation log and status record.

7. The data redundancy remote multi-active system based on industrial Internet of Things according to claim 5 is characterized in that: In the module M2, the process of identifying the key data is to select data and determine whether the value of the selected data exceeds a preset threshold. If the result is yes, the selected data is classified as key data; if the result is no, the selected data is classified as ordinary data; In the module M2, the key data is enhanced by an AI prediction module; the AI ​​prediction module is used to analyze input data and generate prediction results; The key data is composed of features; the mathematical expression of the features is: in, Indicates Features represents the normalization function for feature extraction; Indicates Key data, Indicates the first time offset; Indicates Key data, Indicates the second time offset; In the module M2, the AI ​​prediction module is used for data prediction, and the mathematical expression of the data prediction is: =σ(W +b) in, represents the probability of failure, W represents the weight, b represents the bias vector, and σ represents the activation function; the activation function is Sigmoid or Softmax; Based on the data prediction, the optimization of the fault switching time of the AI ​​prediction module is expressed as follows: in, Indicates the optimization of failover time, The time for the prediction model to respond in advance. is the data redundancy time, The failover time.

8. The data redundancy remote multi-active system based on industrial Internet of Things according to claim 5 is characterized in that: In the module M3, the early warning signal is a warning of high risk or sudden increase in load; The process of judging the high risk or load surge is to judge whether the current CPU or memory usage is greater than 80%, the data throughput per unit time is greater than 10MB / s, or the failure probability score of the AI ​​prediction model of the edge node is greater than 0.

75. If the result is yes, a high risk or load surge warning is issued, and the number of copies of the key data is increased to 3-5 copies, and the synchronization cycle is shortened to 5-10 seconds; If the result is no, the original data redundancy strategy and synchronization cycle are maintained, and detection continues.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the data redundancy remote multi-active method based on the industrial Internet of Things described in any one of claims 1 to 4 are implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is executed by a processor, the steps of the data redundancy remote multi-active method based on the industrial Internet of Things described in any one of claims 1 to 4 are implemented.

Images